What the data actually shows
The most cited finding comes from Przybylski and Weinstein (2013), who had pairs of strangers talk with either a phone or a neutral object nearby. Conversations held in the presence of a phone — even one no one touched — were rated lower on closeness, connection, and conversation quality, with the effect strongest for more meaningful conversations. The phone didn't need to ring or be used; its visible presence alone was associated with the dip.
Roberts and David (2016) coined "partner phubbing" and found, in survey research, that people who reported their romantic partner frequently checking a phone during their time together also reported lower relationship satisfaction — a link that ran partly through conflict over phone use and lower-quality time. As survey data, it shows association rather than proving direction, but it fits the broader pattern.
Across this literature, the proposed mechanism is divided attention and the signal it sends. A phone, present or checked, suggests that part of your attention is reserved for something elsewhere — which people seem to register as reduced engagement even when nothing is said. Importantly, the average effects are modest and not universal; some studies find weaker results, and context (a quick necessary check versus repeated scrolling) clearly matters.
Why this feels different from how it actually is
It feels harmless from your side because you know your own intent — a glance to check the time, a quick reply, nothing personal. But the other person can't see your intent; they see attention leaving. The gap between how a phone check feels to the checker and how it reads to the person across the table is exactly where the social cost lives.
The visible-phone finding is especially counterintuitive because it feels absurd that an object you aren't even touching could matter. The proposed explanation is that a phone is a cue to a whole world of other people and possibilities, and its mere presence subtly signals that this conversation is competing with everything else you could be attending to — a signal that operates below the level you'd consciously endorse.
It also feels different because phone-checking has become so normal that it barely registers as a choice. When everyone does it, the small cost it carries is easy to discount — but the research suggests the cost doesn't disappear just because the behaviour is common; it's simply paid by both people at once and rarely named.
The mere presence of the device seems to shift how present people feel to one another, often below conscious awareness.
What the research says to do about it
The simplest evidence-aligned move is to remove the phone from view during conversations that matter to you, not just to refrain from checking it. Since the visible-presence effect shows up even with an untouched phone, putting it away — pocket, bag, another room — addresses the cue itself rather than only your behaviour.
Where a check is genuinely necessary, briefly naming it appears to help, because much of the cost is the unexplained withdrawal of attention. A short "sorry, I need to reply to this one thing" restores the other person's sense of why your attention left, which the silent glance does not.
Beyond technique, the underlying finding is a reminder that perceived attention is what people respond to. The research on connection consistently points to feeling attended-to as a core ingredient of good interaction, so the practical aim isn't phone-purism — it's protecting the impression, and the reality, that the person in front of you has your attention.
What the research says does not help
Assuming that because everyone does it, it carries no cost, doesn't hold up — the studies suggest the dip in connection appears regardless of how normal the behaviour has become. Universality changes how acceptable phubbing feels, not whether it's associated with lower-quality interaction.
Relying on a quick, silent glance as the "polite" version is also weak, because the cost is the visible withdrawal of attention, which a silent glance still delivers. The other person registers the attention leaving whether or not you say anything.
Treating this as a strict rule to enforce on others tends to backfire and isn't what the data supports. The effects are modest averages, not laws, and policing a partner's or friend's phone use can itself become the source of conflict that the research links to lower satisfaction. The findings are better used as a gentle prompt for your own behaviour than as a charge to level at someone else's.
The cost doesn't disappear just because the behaviour is common; it's simply paid by both people at once and rarely named.
What this looks like in real life
Your intent versus what they see
It feels harmless from your side because you know your own intent — a glance to check the time, a quick reply, nothing personal. But the other person can't see your intent; they see attention leaving. The gap between how a phone check feels to the checker and how it reads to the person across the table is exactly where the social cost lives.
The phone face-down on the table
You set your phone on the table, face-down, and never touch it through dinner. In Przybylski and Weinstein's experiments, even a phone merely visible nearby was associated with lower ratings of closeness and conversation quality. Putting it away — pocket, bag, another room — addresses the cue itself, not just your behaviour.
Real numbers in context
There isn't a single headline percentage here, and the responsible framing is that these are modest, replicated-in-pattern effects rather than dramatic ones. In Przybylski and Weinstein's experiments, the presence of a phone was associated with lower ratings of closeness and conversation quality, with the effect more pronounced for deeper conversations — a small but notable shift produced by an object no one touched.
In survey research like Roberts and David's on "partner phubbing," frequent phone-checking during shared time was associated with lower relationship satisfaction, largely via phone-related conflict and lower-quality time together. Because it's correlational, it can't establish that phubbing causes dissatisfaction rather than the reverse — but it's consistent with the experimental work suggesting that divided attention, signalled by a phone, has a real if modest social cost.